Basal Slices Research Articles

Automated left and right ventricular chamber segmentation in cardiac magnetic resonance images using dense fully convolutional neural network

Background and objectiveSegmentation of the left ventricular (LV) myocardium (Myo) and RV endocardium on cine cardiac magnetic resonance (CMR) images represents an essential step for cardiac-function evaluation and diagnosis. In order to have a common reference for comparing segmentation algorithms, several CMR image datasets were made available, but in general they do not include the most apical and basal slices, and/or gold standard tracing is limited to only one of the two ventricles, thus not fully corresponding to real clinical practice. Our aim was to develop a deep learning (DL) approach for automated segmentation of both RV and LV chambers from short-axis (SAX) CMR images, reporting separately the performance for basal slices, together with the applied criterion of choice. MethodA retrospectively selected database (DB1) of 210 cine sequences (3 pathology groups) was considered: images (GE, 1.5 T) were acquired at Centro Cardiologico Monzino (Milan, Italy), and end-diastolic (ED) and end-systolic frames (ES) were manually segmented (gold standard, GS). Automatic ED and ES RV and LV segmentation were performed with a U-Net inspired architecture, where skip connections were redesigned introducing dense blocks to alleviate the semantic gap between the U-Net encoder and decoder. The proposed architecture was trained including: A) the basal slices where the Myo surrounded the LV for at least the 50% and all the other slice; B) all the slices where the Myo completely surrounded the LV. To evaluate the clinical relevance of the proposed architecture in a practical use case scenario, a graphical user interface was developed to allow clinicians to revise, and correct when needed, the automatic segmentation. Additionally, to assess generalizability, analysis of CMR images obtained in 12 healthy volunteers (DB2) with different equipment (Siemens, 3T) and settings was performed. ResultsThe proposed architecture outperformed the original U-Net. Comparing the performance on DB1 between the two criteria, no significant differences were measured when considering all slices together, but were present when only basal slices were examined. Automatic and manually-adjusted segmentation performed similarly compared to the GS (bias±95%LoA): LVEDV -1±12 ml, LVESV -1±14 ml, RVEDV 6±12 ml, RVESV 6±14 ml, ED LV mass 6±26 g, ES LV mass 5±26 g). Also, generalizability showed very similar performance, with Dice scores of 0.944 (LV), 0.908 (RV) and 0.852 (Myo) on DB1, and 0.940 (LV), 0.880 (RV), and 0.856 (Myo) on DB2. ConclusionsOur results support the potential of DL methods for accurate LV and RV contours segmentation and the advantages of dense skip connections in alleviating the semantic gap generated when high level features are concatenated with lower level feature. The evaluation on our dataset, considering separately the performance on basal and apical slices, reveals the potential of DL approaches for fast, accurate and reliable automated cardiac segmentation in a real clinical setting.

Global, segmental and layer specific analysis of myocardial involvement in Duchenne muscular dystrophy by cardiovascular magnetic resonance native T1 mapping

BackgroundProgressive cardiomyopathy accounts for almost all mortality among Duchenne muscular dystrophy (DMD) patients.‍ Thus, our aim was to comprehensively characterize myocardial involvement by investigating the heterogeneity of native T1 mapping in DMD patients using global and regional (including segmental and layer-specific) analysis across a large cohort.MethodsWe prospectively enrolled 99 DMD patients (8.8 ± 2.5 years) and 25 matched male healthy controls (9.5 ± 2.5 years). All subjects underwent cardiovascular magnetic resonance (CMR) with cine, T1 mapping and late gadolinium enhancement (LGE) sequences. Native T1 values based on the global and regional myocardium were measured, and LGE was defined.ResultsLGE was present in 49 (49%) DMD patients. Global native T1 values were significantly longer in LGE-positive (LGE +) patients than in healthy controls, both in basal slices (1304 ± 55 vs. 1246 ± 27 ms, p < 0.001) and in mid-level slices (1305 ± 57 vs. 1245 ± 37 ms, p < 0.001). No significant difference in global native T1 was found between healthy controls and LGE-negative (LGE−) patients. In segmental analysis, LGE + patients had significantly increased native T1 in all analyzed segments compared to the healthy control group. Meanwhile, the comparison between LGE− patients and healthy controls showed significantly elevated values only in the basal anterolateral segment (1273 ± 62 vs. 1234 ± 40 ms, p = 0.034). Interestingly, the epicardial layer had a significantly higher native T1 in LGE− patients than in healthy controls (p < 0.05), whereas no such pattern was noticed in the global myocardium. Epicardial layer native T1 resulted in the highest diagnostic performance for distinguishing between healthy controls and DMD patients in receiver operating curve analyses (area under the curve [AUC] 0.84 for basal level and 0.85 for middle level) when compared to global native T1 and endocardial layer native T1.ConclusionsMyocardial regional native T1, particularly epicardial native T1, seems to have potential as a novel robust marker of very early cardiac involvement in DMD patients.Trial registration: Chinese Clinical Trial Registry (http://www.chictr.org.cn/index.aspx) ChiCTR1800018340, 09/12/2018, Retrospectively registered.

Endogenous T1\u03c1 cardiovascular magnetic resonance in hypertrophic cardiomyopathy

BackgroundHypertrophic cardiomyopathy (HCM) is characterized by increased left ventricular wall thickness, cardiomyocyte hypertrophy, and fibrosis. Adverse cardiac risk characterization has been performed using late gadolinium enhancement (LGE), native T1, and extracellular volume (ECV). Relaxation time constants are affected by background field inhomogeneity. T1ρ utilizes a spin-lock pulse to decrease the effect of unwanted relaxation. The objective of this study was to study T1ρ as compared to T1, ECV, and LGE in HCM patients.MethodsHCM patients were recruited as part of the Novel Markers of Prognosis in Hypertrophic Cardiomyopathy study, and healthy controls were matched for comparison. In addition to cardiac functional imaging, subjects underwent T1 and T1ρ cardiovascular magnetic resonance imaging at short-axis positions at 1.5T. Subjects received gadolinium and underwent LGE imaging 15–20 min after injection covering the entire heart. Corresponding basal and mid short axis LGE slices were selected for comparison with T1 and T1ρ. Full-width half-maximum thresholding was used to determine the percent enhancement area in each LGE-positive slice by LGE, T1, and T1ρ. Two clinicians independently reviewed LGE images for presence or absence of enhancement. If in agreement, the image was labeled positive (LGE + +) or negative (LGE −−); otherwise, the image was labeled equivocal (LGE + −).ResultsIn 40 HCM patients and 10 controls, T1 percent enhancement area (Spearman’s rho = 0.61, p < 1e-5) and T1ρ percent enhancement area (Spearman’s rho = 0.48, p < 0.001e-3) correlated with LGE percent enhancement area. T1 and T1ρ percent enhancement areas were also correlated (Spearman’s rho = 0.28, p = 0.047). For both T1 and T1ρ, HCM patients demonstrated significantly longer relaxation times compared to controls in each LGE category (p < 0.001 for all). HCM patients also showed significantly higher ECV compared to controls in each LGE category (p < 0.01 for all), and LGE −− slices had lower ECV than LGE + + (p = 0.01).ConclusionsHyperenhancement areas as measured by T1ρ and LGE are moderately correlated. T1, T1ρ, and ECV were elevated in HCM patients compared to controls, irrespective of the presence of LGE. These findings warrant additional studies to investigate the prognostic utility of T1ρ imaging in the evaluation of HCM patients.

Identifying Myocardial Infarction Using Hierarchical Template Matching-Based Myocardial Strain: Algorithm Development and Usability Study.

BackgroundMyocardial infarction (MI; location and extent of infarction) can be determined by late enhancement cardiac magnetic resonance (CMR) imaging, which requires the injection of a potentially harmful gadolinium-based contrast agent (GBCA). Alternatively, emerging research in the area of myocardial strain has shown potential to identify MI using strain values.ObjectiveThis study aims to identify the location of MI by developing an applied algorithmic method of circumferential strain (CS) values, which are derived through a novel hierarchical template matching (HTM) method.MethodsHTM-based CS H-spread from end-diastole to end-systole was used to develop an applied method. Grid-tagging magnetic resonance imaging was used to calculate strain values in the left ventricular (LV) myocardium, followed by the 16-segment American Heart Association model. The data set was used with k-fold cross-validation to estimate the percentage reduction of H-spread among infarcted and noninfarcted LV segments. A total of 43 participants (38 MI and 5 healthy) who underwent CMR imaging were retrospectively selected. Infarcted segments detected by using this method were validated by comparison with late enhancement CMR, and the diagnostic performance of the applied algorithmic method was evaluated with a receiver operating characteristic curve test.ResultsThe H-spread of the CS was reduced in infarcted segments compared with noninfarcted segments of the LV. The reductions were 30% in basal segments, 30% in midventricular segments, and 20% in apical LV segments. The diagnostic accuracy of detection, using the reported method, was represented by area under the curve values, which were 0.85, 0.82, and 0.87 for basal, midventricular, and apical slices, respectively, demonstrating good agreement with the late-gadolinium enhancement–based detections.ConclusionsThe proposed applied algorithmic method has the potential to accurately identify the location of infarcted LV segments without the administration of late-gadolinium enhancement. Such an approach adds the potential to safely identify MI, potentially reduce patient scanning time, and extend the utility of CMR in patients who are contraindicated for the use of GBCA.

Tissue characteristics of the athlete"s heart: differentiation of physiological and pathological hypertrophy using parametric T1 and T2 mapping

Abstract Funding Acknowledgements Type of funding sources: Public grant(s) – National budget only. Main funding source(s): National Research, Development and Innovation Fund of Hungary Background Intensive physical exercise leads to structural and functional cardiac adaptation termed athlete’s heart. Cardiac magnetic resonance (CMR) has an important role in the differentiation of physiological adaptation and pathological conditions. Beside the precise measurement of the ventricular volumes, mass, and function, it provides tissue specific information. Recently, native T1 mapping technique has been applied as a non-contrast method to detect myocardial fibrosis. Previous studies suggested that native T1 mapping can identify myocardial pathology before other CMR imaging techniques. T2 mapping values are elevated in case of myocardial edema. Purpose The aim of our study was to investigate the differences in CMR characteristics especially the native T1 and T2 mapping values of highly trained healthy athletes, healthy controls and patients with hypertrophic cardiomyopathy (HCM). Methods A total of 43 healthy athletes (water polo, swimming, football, 22 ± 8 training hours/week), 27 non-athlete healthy control and 25 HCM patients were involved in the study. Our inclusion criteria were: age &amp;gt;18 years, in the athlete group &amp;gt;7 training hours per week . We evaluated the left ventricular (LV) end-systolic, end-diastolic (EDVi) and stroke volume (SVi) index, mass index (LVMi), ejection fraction (EF) and maximal end-diastolic wall thickness (EDWT). In a basal short axis slice the native T1 and T2 mapping values were evaluated. Results Athletes had significantly higher LV volumes compared to the control and HCM group (LVEDVi 114 ± 13 vs. 86 ± 11; 84 ± 15 ml/m2, LVSVi 64 ± 7 vs. 51 ± 7; 54 ± 10 ml/m2, respectively, p &amp;lt; 0.0001). HCM patients had the highest LVMi (72 ± 14 g/m2) and EDWT (18 ± 4 mm) compared to athletes and controls, athletes had higher LVMi (60 ± 11 vs. 42 ± 8 g/m2) and EDWT (10 ± 2 vs. 8 ± 1 mm) compared to the controls (p &amp;lt; 0.001). The native T1 mapping values differed significantly in the three groups, athletes had the lowest, HCM patients had the highest T1 values (athletes: 956 ± 19 ms, controls: 971 ± 20 ms, HCM patients: 993 ± 39 ms; p &amp;lt; 0.0001). There was no difference in the T2 mapping values between athletes and controls (44 ± 2 vs. 43 ± 2 ms), HCM patients had higher T2 values (45 ± 2 ms) compared to the other two groups (p &amp;lt; 0.01). Conclusion Intensive and regular training may lead to tissue specific changes of the myocardium. T1 and T2 mapping are potentially useful tools for differentiating between athlete"s heart and patients with hypertrophic cardiomyopathy. Abstract Figure. T1 mapping in HCM and athlete

Evolution of biventricular T1 values in patients with right-sided congenital heart disease

Abstract Funding Acknowledgements Type of funding sources: None. Background Conflicting reports exist on the prevalence and clinical impact of interstitial fibrosis in right ventricular (RV) congenital heart disease (CHD). This study evaluates the longitudinal evolution of native myocardial T1 relaxation time (T1) in RV CHD. Methods On a 1.5T scanner, an ECG-triggered modified Look-Locker inversion recovery sequence (scheme 3(3)3(3)5) was acquired on a short-axis basal slice covering the RV and left ventricle (LV) on two consecutive CMR exams. Global and segmental (LV = 6, RV = 4) RV and LV T1 values were calculated (Figure). Results Mean time between CMR exams for 36 included patients (age 34 ± 2y) was 22 ± 2 months. All LV segments and 81/88% of RV segments of first and second CMR could be analyzed, respectively. T1 increased mildly but not significantly (table). There was no relationship of T1 to pulmonary regurgitation fraction, pulmonary stenosis or RV enddiastolic volume (p &amp;gt; 0.05). Global RV T1 of the second CMR was related to RV ejection fraction (RVEF): r = 0.353, 3.0 ± 1.4, p = 0.038. T1 of the infero-septal LV segment of first and second CMR, global LV T1 of second CMR and increase of T1 of global LV, anterior, antero-lateral and –septal LV segments, were related to age at CMR: r = 0.333 - 0.463, p &amp;lt; 0.05, respectively. Conclusions Native T1 values increased mildly in patients with stable RV CHD, which was not statistically significant probably due to the short to median follow-up. Global RV T1 appears to be related to RVEF which could be sign of increasing interstitial fibrosis whereas the relationship of LV T1 to age might be a physiological finding. First CMR native T1 (ms) Second CMR native T1 (ms) p LV Global 1007 ± 37 1014 ± 39 0.413 LV Anterior 994 ± 53 999 ± 54 0.710 LV Antero-lateral 965 ± 63 981 ± 58 0.186 LV Infero-lateral 1000 ± 52 1004 ± 63 0.695 LV Inferior 1035 42 1037 ± 50 0.744 LV Infero-septal 1028 ± 35 1036 ± 43 0.282 LV Antero-septal 1016 ± 38 1024 ± 48 0.347 RV Global 1091 ± 90 1096 ± 85 0.410 RV Inferior 1112 ± 104 1115 ± 118 0.696 RV Infero-lateral 1061 ± 130 1077 ± 115 0.425 RV Antero-lateral 1046 ± 127 1080 ± 109 0.088 RV Anterior 1088 ± 156 1108 ± 154 0.410 Abstract Figure. Determination of biventricular T1 values

Networking of glucagon-like peptide-1 axons with GnRH neurons in the basal forebrain of male mice revealed by 3DISCO-based immunocytochemistry and optogenetics

Glucagon-like peptide-1 (GLP-1) regulates reproduction centrally, although, the neuroanatomical basis of the process is unknown. Therefore, the putative networking of the central GLP-1 and gonadotropin-releasing hormone (GnRH) systems was addressed in male mice using whole mount immunocytochemistry and optogenetics. Enhanced antibody penetration and optical clearing procedures applied to 500–1000 µm thick basal forebrain slices allowed the simultaneous visualization of the two distinct systems in the basal forebrain. Beaded GLP-1-IR axons innervated about a quarter of GnRH neurons (23.2 ± 1.4%) forming either single or multiple contacts. GnRH dendrites received a more intense GLP-1 innervation (64.6 ± 0.03%) than perikarya (35.4 ± 0.03%). The physiological significance of the innervation was examined by optogenetic activation of channelrhodopsin-2 (ChR2)-expressing axons of preproglucagon (GCG) neurons upon the firing of GnRH neurons by patch clamp electrophysiology in acute brain slices of triple transgenic mice (Gcg-cre/ChR2/GFP-GnRH). High-frequency laser beam stimulation (20 Hz, 10 ms pulse width, 3 mW laser power) of ChR2-expressing GCG axons in the mPOA increased the firing rate of GnRH neurons (by 75 ± 17.3%, p = 0.0007). Application of the GLP-1 receptor antagonist, Exendin-3-(9-39) (1 μM), prior to the photo-stimulation, abolished the facilitatory effect. In contrast, low-frequency trains of laser pulses (0.2 Hz, 60 pulses) had no effect on the spontaneous postsynaptic currents of GnRH neurons. The findings indicate a direct wiring of GLP-1 neurons with GnRH cells which route is excitatory for the GnRH system. The pathway may relay metabolic signals to GnRH neurons and synchronize metabolism with reproduction.

Texture-based machine learning in LGE-CMRI for automatic scar assessment

Abstract Background Conventional methods for the estimation of infarct size by late-enhanced cardiac magnetic resonance imaging use analyzes from each slice which in turn is added to generate a volume. We present a novel method based on machine learning, called texture based probability mapping (TPM). TPM is based upon expressing the probability of scarring in a pixel based upon analysing textural information in patches around each pixel. Purpose To explore our method and assess its utility as a tool to scar size and compare it with four established methods. Methods In 54 patients with ischemic scars, our method was compared with four 2D methods. 2 patients were excluded due to low image quality and artefacts. The most basal and apical short axis image slices were exclude due to partial volume artefacts. Myocardial infarction (MI) size was estimated manually, and TPM performed. Cardiac segments were defined as myocardial regions with pixel probabilities within a specified range. By varying the lower probability threshold while keeping the upper threshold fixed at 1, different cardiac segments were defined. Using the Sørensen-Dice coefficient the optimal probability range with the highest correlation to manual estimation of infarct size was identified. Results TPM-range 0.341–1.0 is best correlated to manual demarcation (median Dice 0.70). The method demonstrated stronger correlations between scar size and left ventricular remodeling parameters (LV ejection fraction: r=−0.731, p&amp;lt;0.0005; LV end-diastolic volume: r=0.641, p&amp;lt;0.0005; LV end-systolic volume: r=0.672, p&amp;lt;0.0005) compared with manual method. Conclusion Using TPM, infarct size can be measured automatically without using signal intensity as a reference value. It is without the need for manual demarcation of the scar and is less time consuming and less vulnerable to interobserver variability. It has a strong correlation with left ventricular remodeling parameters crucial in guiding further medical therapy. The use of this method is however not primarily dedicated for scar size estimation, but a tool to evaluate different properties of the tissue that cannot be visualized by the naked eye. It's an important step towards better understanding of myocardial damage pattern causing ventricular arrhythmia and SCD. Funding Acknowledgement Type of funding source: Public grant(s) – National budget only. Main funding source(s): Helse Vest

Structural and functional cardiac changes after transplantation: experiences of the first year of the prospective Heart-TIming CMR substudy

Abstract Background In case of heart transplantation (HTX) the heart is affected by several factors e.g. ischaemia/reperfusion, denervation, immunosuppression. During the adaptation, the heart may show marked temporal changes in terms of myocardial mechanics, function and tissue characteristics. To better understand cardiac temporal characteristics after orthotopic bicaval HTX we started the prospective Heart-TIming (Transplantation Imaging) trial in January 2018. Purpose In our CMR substudy we aimed to evaluate the physiological structural and functional left and right ventricular characteristics and their temporal changes after HTX using cardiac magnetic resonance. Methods As part of the study HTX patients underwent CMR at one, three, six and twelve months after HTX (n=49; 53±11y, 39 male). Cine images, T2-weighted, native T1 and T2 mapping, late gadolinium enhancement (LGE) and adenosine stress perfusion (at 1 and 12 month) images were acquired. In order to describe physiological characteristics of the transplanted heart we excluded pts with significant coronary artery disease, ischaemic scar, ≥Grade II allograft rejection from this present study (n=9). We evaluated the left (LV) and right ventricular (RV) ejection fractions (EF), volumes, masses (M) and the global LV strain values: longitudinal (GLS), circumferential (GCS) strain and the standard deviation (SD) of the peak longitudinal strain (LS) and the mechanical dispersion. In a basal short axis slice the native T1 and T2 mapping values were evaluated. We compared baseline CMR parameters to age and gender matched healthy controls (n=20; 48±10y, 16 male), and analyzed the temporal changes after HTX. Results Comparing the HTX patients' CMR parameters at one month with normal controls, HTX patients had lower end-diastolic volumes (LVEDVi: 74±15 vs 89±13 ml/m2; RVEDVi: 72±16 vs 89±15 ml/m2 p&amp;lt;0.05), stroke volumes (LVSVi: 45±7 vs 55±8 ml/m2, RVSVi: 43±8 vs 54±8 ml/m2, p&amp;lt;0.0001), higher LVMi (63±2 vs 55±3 g/m2, p&amp;lt;0.05), increased SD of peak LS (14±2 vs 10±2, p&amp;lt;0.0001) and more pronounced mechanical dispersion (18±5 vs 12±4, p&amp;lt;0.0001). The native T1 mapping values were significantly higher in HTX pts (1007±40 vs 975±24 ms, p&amp;lt;0,01). Examining temporal changes in HTX pts we found a decrease in LVMi (66±14 vs 59±10 g/m2, p&amp;lt;0.01) already at three months. At 12 months LVMi decreased further, less negative GLS (−25±4 vs −20±4, p&amp;lt;0.01) and GCS (−38±7 vs −34±4, p&amp;lt;0.05), and lower SD of the peak LS (14±2 vs 11±2, p&amp;lt;0.01) were measured. Conclusions Understanding the temporal changes of cardiac mechanics, function and tissue characteristics, furthermore the establishment of physiological values may help in the early, noninvasive identification of pathological changes in HTX pts. Tissue specific information in HTX pts Funding Acknowledgement Type of funding source: Public grant(s) – National budget only. Main funding source(s): Project no. NVKP_16-1-2016-0017 has been implemented with the support provided from the National Research, Development and Innovation Fund of Hungary, financed under the NVKP_16 funding scheme. Supported by the ÚNKP-18-3-IV New National Excellence Program of Human Capacities.

Comparability of compressed sensing-based gradient echo perfusion sequence SPARSE and conventional gradient echo sequence in assessment of myocardial ischemia

PurposeStress perfusion imaging plays a major role in non-invasive detection of coronary artery disease.We compared a compressed sensing-based and a conventional gradient echo perfusion sequence with regard to image quality and diagnostic performance. MethodPatients sent for coronary angiography due to pathologic stress perfusion CMR were recruited. All patients underwent two adenosine stress CMR using conventional TurboFLASH and prototype SPARSE sequence as well as quantitative coronary angiography with fractional flow reserve (FFR) within 6 weeks. Coronary angiography was considered gold standard with FFR < 0.75 or visual stenosis >90 % for identification of myocardial ischemia. Diagnostic performance of perfusion imaging was assessed in basal, mid-ventricular and apical slices by quantification of myocardial perfusion reserve (MPR) analysis utilizing the signal upslope method and a deconvolution technique using the fermi function model. Results23 patients with mean age of 69.6 ± 8.9 years were enrolled. 46 % were female.Image quality was similar in conventional TurboFLASH sequence and SPARSE sequence (2.9 ± 0.5 vs 3.1 ± 0.7, p = 0,06). SPARSE sequence showed higher contrast-to-noise ratio (52.1 ± 27.4 vs 40.5 ± 17.6, p < 0.01) and signal-to-noise ratio (15.6 ± 6.2 vs 13.2 ± 4.2, p < 0.01) than TurboFLASH sequence. Dark-rim artifacts occurred less often with SPARSE (9 % of segments) than with TurboFLASH (23 %).In visual assessment of perfusion defects, SPARSE sequence detected less false-positive perfusion defects (n = 1) than TurboFLASH sequence (n = 3).Quantitative perfusion analysis on segment basis showed equal detection of perfusion defects for TurboFLASH and SPARSE with both upslope MPR analysis (TurboFLASH 0.88 ± 0.18; SPARSE 0.77 ± 0.26; p = 0.06) and fermi function model (TurboFLASH 0.85 ± 0.24; SPARSE 0.76 ± 0.30; p = 0.13). ConclusionsCompressed sensing perfusion imaging using SPARSE sequence allows reliable detection of myocardial ischemia.

Prognostic value of left ventricular remodelling index in idiopathic dilated cardiomyopathy.

To evaluate the prognostic value of left ventricular (LV) remodelling index (RI) in idiopathic dilated cardiomyopathy (DCM) patients. We prospectively enrolled 412 idiopathic DCM patients and 130 age- and sex-matched healthy volunteers who underwent cardiovascular magnetic resonance imaging between September 2013 and March 2018. RI was defined as the cubic root of the LV end-diastolic volume divided by the mean LV wall thickness on basal short-axis slice. The primary endpoint included all-cause mortality and heart transplantation. The secondary endpoint included the primary endpoint and heart failure (HF) readmission. During the median follow-up of 28.1 months (interquartile range: 19.3-43.0 months), 62 (15.0%) and 143 (34.7%) patients reached the primary and secondary endpoints, respectively. Stepwise multivariate Cox regression showed that RI [hazard ratio (HR) 1.20, 95% confidence interval (CI) 1.11-1.30, P < 0.001], late gadolinium enhancement (LGE) presence and log (N-terminal pro-B-type natriuretic peptide) were independent predictors of the primary endpoint, while RI (HR 1.15, 95% CI 1.08-1.23, P < 0.001) and extracellular volume were independent predictors of the secondary endpoint. The addition of RI to LV ejection fraction (EF) and LGE presence showed significantly improved global χ2 for predicting primary and secondary endpoints (both P < 0.001). Furthermore, RI derived from echocardiography also showed independent prognostic value for primary and secondary endpoints with clinical risk factors. RI is an independent predictor of all-cause mortality, heart transplantation, and HF readmission in DCM patients and provides incremental prognostic value to LVEF and LGE presence.

Clinical Performance and Role of Expert Supervision of Deep Learning for Cardiac Ventricular Volumetry: A Validation Study

To evaluate the performance of a deep learning (DL) algorithm for clinical measurement of right and left ventricular volume and function across cardiac MR images obtained for a range of clinical indications and pathologies. A retrospective, Health Insurance Portability and Accountability Act-compliant study was conducted using the first 200 noncongenital clinical cardiac MRI examinations from June 2015 to June 2017 for which volumetry was available. Images were analyzed using commercially available software for automated DL-based and manual contouring of biventricular volumes. Fully automated measurements were compared using Pearson correlations, relative volume errors, and Bland-Altman analyses. Manual, automated, and expert revised contours for 50 MR images were examined by comparing regional Dice coefficients at the base, midventricle, and apex to further analyze the contour quality. Fully automated and manual left ventricular volumes were strongly correlated for end-systolic volume (ESV: Pearson r = 0.99, P < .001), end-diastolic volume (EDV: r = 0.97, P < .001), and ejection fraction (EF: r = 0.94, P < .001). Right ventricular measurements were also correlated for ESV (r = 0.93, P < .001), EDV (r = 0.92, P < .001), and EF (r = 0.73, P < .001). Visual inspection of segmentation quality showed most errors (73%) occurred at the cardiac base. Mean Dice coefficients between manual, automated, and expert revised contours ranged from 0.92 to 0.95, with greatest variance at the base and apex. Fully automated ventricular segmentation by the tested algorithm provides contours and ventricular volumes that could be used to aid expert segmentation, but can benefit from expert supervision, particularly to resolve errors at the basal and apical slices. Supplemental material is available for this article. © RSNA, 2020.

Radiomic Analysis of Native T1 Mapping Images Discriminates Between MYH7 and MYBPC3-Related Hypertrophic Cardiomyopathy.

The phenotype via conventional cardiac MRI analysis of MYH7 (β-myosin heavy chain)- and MYBPC3 (β-myosin-binding protein C)-associated hypertrophic cardiomyopathy (HCM) groups is similar. Few studies exist on the genotypic-phenotypic association as assessed by machine learning in HCM patients. To explore the phenotypic differences based on radiomics analysis of T1 mapping images between MYH7 and MYBPC3-associated HCM subgroups. Prospective observational study. In all, 102 HCM patients with pathogenic, or likely pathogenic mutation, in MYH7 (n = 68) or MYBPC3 (n = 34) genes. Cardiac MRI was performed at 3.0T with balanced steady-state free precession (bSSFP), phase-sensitive inversion recovery (PSIR) late gadolinium enhancement (LGE), and modified Look-Locker inversion recovery (MOLLI) T1 mapping sequences. All patients underwent next-generation sequencing and Sanger genetic sequencing. Left ventricular native T1 and LGE were analyzed. One hundred and fifty-seven radiomic features were extracted and modeled using a support vector machine (SVM) combined with principal component analysis (PCA). Each subgroup was randomly split 4:1 (feature selection / test validation). Mann-Whitney U-tests and Student's t-tests were performed to assess differences between subgroups. A receiver operating characteristic (ROC) curve was used to assess the model's ability to stratify patients based on radiomic features. There were no significant differences between MYH7- and MYBPC3-associated HCM subgroups based on traditional native T1 values (global, basal, and middle short-axis slice native T1 ; P = 0.760, 0.914, and 0.178, respectively). However, the SVM model combined with PCA achieved an accuracy and area under the curve (AUC) of 92.0% and 0.968 (95% confidence interval [CI]: 0.968-0.971), respectively. For the test validation dataset, the accuracy and AUC were 85.5% and 0.886 (95% CI: 0.881-0.901), respectively. Radiomic analysis of native T1 mapping images may be able to discriminate between MYH7- and MYBPC3-associated HCM patients, exceeding the performance of conventional native T1 values. 3 TECHNICAL EFFICACY STAGE: 2 J. MAGN. RESON. IMAGING 2020;52:1714-1721.

Diagnostic Accuracy of Mapping Techniques and Postprocessing Methods for Acute Myocarditis.

OBJECTIVE. Many T1 mapping sequences and measurement methods have been reported in the literature. To obtain accurate results, the most appropriate T1 and T2 mapping sequences and postprocessing methods should be used. The aim of the present study is to compare the T1 and T2 mapping sequences and measurement methods used to diagnose acute myocarditis. MATERIALS AND METHODS. Fourteen patients with an interval of 7 days or less between symptom onset and MRI examination were included in the study. The control group consisted of 22 healthy individuals. T1 mapping images acquired using four different methods (modified Look-Locker inversion recovery [MOLLI] 3[3]3[3]5, MOLLI 3[2]3[2]5, MOLLI 5[3]3 heart rate-corrected, and MOLLI 5[3]3 heart rate-uncorrected sequences) and T2 mapping images obtained using two different methods (FLASH and true fast imaging with steady-state precession) in the area from the short axis through the basal slice were retrospectively evaluated. RESULTS. The mean T1 and T2 values were significantly higher in the patient group than in the control group for all sequences in SAX measurements and segmental measurements (segments 1, 4, 5, and 6). Measurements from segments 2 and 3 were not statistically significant between patient and control groups. The segment with the highest AUC value of T1 and T2 mapping values was the inferior segment, followed by the inferolateral segment. The highest diagnostic performance for short-axis measurements was obtained using the MOLLI 3(2)3(2)5 sequence (AUC value = 0.902). CONCLUSION. According to the results of this study and many other studies, pathologic changes in the myocardium can be detected by native T1 and T2 mapping. The accuracy of mapping methods will increase with increased knowledge of the advantages and disadvantages of currently used sequences and measurement methods.

Simultaneous Mapping of T1 and T2 Using Cardiac Magnetic Resonance Fingerprinting in a Cohort of Healthy Subjects at 1.5T.

Cardiac MR fingerprinting (cMRF) is a novel technique for simultaneous T1 and T2 mapping. To compare T1 /T2 measurements, repeatability, and map quality between cMRF and standard mapping techniques in healthy subjects. Prospective. In all, 58 subjects (ages 18-60). FIELD STRENGTH/SEQUENCE: cMRF, modified Look-Locker inversion recovery (MOLLI), and T2 -prepared balanced steady-state free precession (bSSFP) at 1.5T. T1 /T2 values were measured in 16 myocardial segments at apical, medial, and basal slice positions. Test-retest and intrareader repeatability were assessed for the medial slice. cMRF and conventional mapping sequences were compared using ordinal and two alternative forced choice (2AFC) ratings. Paired t-tests, Bland-Altman analyses, intraclass correlation coefficient (ICC), linear regression, one-way analysis of variance (ANOVA), and binomial tests. Average T1 measurements were: basal 1007.4±96.5 msec (cMRF), 990.0±45.3 msec (MOLLI); medial 995.0±101.7 msec (cMRF), 995.6±59.7 msec (MOLLI); apical 1006.6±111.2 msec (cMRF); and 981.6±87.6 msec (MOLLI). Average T2 measurements were: basal 40.9±7.0 msec (cMRF), 46.1±3.5 msec (bSSFP); medial 41.0±6.4 msec (cMRF), 47.4±4.1 msec (bSSFP); apical 43.5±6.7 msec (cMRF), 48.0±4.0 msec (bSSFP). A statistically significant bias (cMRF T1 larger than MOLLI T1 ) was observed in basal (17.4 msec) and apical (25.0 msec) slices. For T2 , a statistically significant bias (cMRF lower than bSSFP) was observed for basal (-5.2 msec), medial (-6.3 msec), and apical (-4.5 msec) slices. Precision was lower for cMRF-the average of the standard deviation measured within each slice was 102 msec for cMRF vs. 61 msec for MOLLI T1 , and 6.4 msec for cMRF vs. 4.0 msec for bSSFP T2 . cMRF and conventional techniques had similar test-retest repeatability as quantified by ICC (0.87 cMRF vs. 0.84 MOLLI for T1 ; 0.85 cMRF vs. 0.85 bSSFP for T2 ). In the ordinal image quality comparison, cMRF maps scored higher than conventional sequences for both T1 (all five features) and T2 (four features). This work reports on myocardial T1 /T2 measurements in healthy subjects using cMRF and standard mapping sequences. cMRF had slightly lower precision, similar test-retest and intrareader repeatability, and higher scores for map quality. 2 TECHNICAL EFFICACY: Stage 1 J. Magn. Reson. Imaging 2020;52:1044-1052.

Longitudinal assessment of tissue properties and cardiac diffusion metrics of the ex vivo porcine heart at 7 T: Impact of continuous tissue fixation using formalin

In this study we aimed to assess the effects of continuous formalin fixation on diffusion and relaxation metrics of the ex vivo porcine heart at 7 T. Magnetic resonance imaging was performed on eight piglet hearts using a 7 T whole body system. Hearts were measured fresh within 3 hours of cardiac arrest followed by immersion in 10% neutral buffered formalin. T2 * and T2 were assessed using a gradient multi-echo and multi-echo spin echo sequence, respectively. A spin echo and a custom stimulated echo sequence were employed to assess diffusion time-dependent changes in metrics of cardiac diffusion tensor imaging. SNR was determined for b = 0 images. Scans were performed for 5 mm thick apical, midcavity and basal slices (in-plane resolution: 1 mm) and repeated 7, 15, 50, 100 and 200 days postfixation. Eigenvalues of the apparent diffusion coefficient (ADC) and fractional anisotropy (FA) decreased significantly (P < 0.05) following fixation. Relative to fresh hearts, FA values 7 and 200 days postfixation were 90% and 80%, while respective relative ADC values at those fixation stages were 78% and 92%. Statistical helix and sheetlet angle distributions as well as respective mean and median values showed no systematic influence of continuous formalin fixation. Similar to changes in the ADC, values for T2 , T2 * and SNR dropped initially postfixation. Respective relative values compared with fresh hearts at day 7 were 64%, 79% and 68%, whereas continuous fixation restored T2 , T2 * and SNR leading to relative values of 74%, 100%, and 81% at day 200, respectively. Relaxation parameters and diffusion metrics are significantly altered by continuous formalin fixation. The preservation of microstructure metrics following prolonged fixation is a key finding that may enable future studies of ventricular remodeling in cardiac pathologies.

Single-shot EPI for ASL-CMR.

To evaluate single-shot echo planar imaging (SS-EPI), as an alternative to snapshot balanced steady state free precession (bSSFP) imaging, for arterial-spin-labeled cardiac MR (ASL-CMR). This study presents a practical implementation SS-EPI tailored to the needs of ASL-CMR at 3T and demonstrates sequential multi-slice ASL with no increase in scan time. Reduced field of view SS-EPI was performed using a 2DRF pulse. A spin-echo was used with crushers optimized to maximize blood suppression and minimize myocardial signal loss, based on experiments in 4 healthy volunteers. SS-EPI was evaluated against the widely used bSSFP reference method in single-slice ASL-CMR in 4 healthy volunteers, during both systole and diastole. Sequential multi-slice ASL-CMR with SS-EPI was demonstrated during diastole (3 slices: basal, mid, and apical short-axis) and during systole (2 slices: mid and apical short-axis), in 3 volunteers. Global myocardial perfusion for diastolic SS-EPI (1.66 ± 0.73 mL/g/min) and systolic SS-EPI (1.50 ± 0.36 mL/g/min) were found to be statistically equivalent (2 one-sided test with a difference of 0.4 mL/g/min) to diastolic bSSFP (duration of 1 cardiac cycle, 1.60 ± 0.80 mL/g/min) with P-values of 0.022 and 0.031, respectively. Global myocardial perfusion for sequential multi-slice experiments was 1.64 ± 0.47, 1.34 ± 0.29, and 1.88 ± 0.58 for basal, mid, and apical SAX slices during diastole and was 1.61 ± 0.35, and 1.66 ± 0.49 for mid and apical slice during systole. These values are comparable to published ASL-CMR and positron emission tomography studies. SS-EPI is a promising alternative to bSSFP imaging for ASL-CMR and can potentially improve the spatial coverage of ASL-CMR by 3-fold during diastole and 2-fold during systole, without increasing scan time.

Large-scale Quality Control of Cardiac Imaging in Population Studies: Application to UK Biobank

In large population studies such as the UK Biobank (UKBB), quality control of the acquired images by visual assessment is unfeasible. In this paper, we apply a recently developed fully-automated quality control pipeline for cardiac MR (CMR) images to the first 19,265 short-axis (SA) cine stacks from the UKBB. We present the results for the three estimated quality metrics (heart coverage, inter-slice motion and image contrast in the cardiac region) as well as their potential associations with factors including acquisition details and subject-related phenotypes. Up to 14.2% of the analysed SA stacks had sub-optimal coverage (i.e. missing basal and/or apical slices), however most of them were limited to the first year of acquisition. Up to 16% of the stacks were affected by noticeable inter-slice motion (i.e. average inter-slice misalignment greater than 3.4 mm). Inter-slice motion was positively correlated with weight and body surface area. Only 2.1% of the stacks had an average end-diastolic cardiac image contrast below 30% of the dynamic range. These findings will be highly valuable for both the scientists involved in UKBB CMR acquisition and for the ones who use the dataset for research purposes.

Fully Automated Myocardial Strain Estimation from Cardiovascular MRI-tagged Images Using a Deep Learning Framework in the UK Biobank.

PurposeTo demonstrate the feasibility and performance of a fully automated deep learning framework to estimate myocardial strain from short-axis cardiac MRI–tagged images.Materials and MethodsIn this retrospective cross-sectional study, 4508 cases from the U.K. Biobank were split randomly into 3244 training cases, 812 validation cases, and 452 test cases. Ground truth myocardial landmarks were defined and tracked by manual initialization and correction of deformable image registration using previously validated software with five readers. The fully automatic framework consisted of (a) a convolutional neural network (CNN) for localization and (b) a combination of a recurrent neural network (RNN) and a CNN to detect and track the myocardial landmarks through the image sequence for each slice. Radial and circumferential strain were then calculated from the motion of the landmarks and averaged on a slice basis.ResultsWithin the test set, myocardial end-systolic circumferential Green strain errors were −0.001 ± 0.025, −0.001 ± 0.021, and 0.004 ± 0.035 in the basal, mid-, and apical slices, respectively (mean ± standard deviation of differences between predicted and manual strain). The framework reproduced significant reductions in circumferential strain in participants with diabetes, hypertensive participants, and participants with a previous heart attack. Typical processing time was approximately 260 frames (approximately 13 slices) per second on a GPU with 12 GB RAM compared with 6–8 minutes per slice for the manual analysis.ConclusionThe fully automated combined RNN and CNN framework for analysis of myocardial strain enabled unbiased strain evaluation in a high-throughput workflow, with similar ability to distinguish impairment due to diabetes, hypertension, and previous heart attack.Published under a CC BY 4.0 license.Supplemental material is available for this article.

1047 Increased interstitial fibrosis in patients with mitral valve prolapse and mitro-annular dysjunction

Abstract Background in patients with myxomatous mitral valve prolapse (MVP), mitral annular disjunction (MAD) has been associated with the presence of late gadolinium enhancement (LGE) at papillary muscle level and the risk of sudden cardiac death. However, patients with MAD but no detectable LGE still may have arrhythmia. We investigated the relation between MAD and the presence of interstitial fibrosis in the basal inferior left ventricular myocardium. Methods 28 patient with MVP and associated MAD underwent Cardiovascular Magnetic Resonance imaging (CMR) at 1.5 T scanner (Aera, Siemens Medical Solutions, Erlangen, Germany). Exclusion criteria were ischemic heart disease, infiltrative cardiomyopathy and contraindication to CMR. 12 patients with mitral valve regurgitation but no MAD and 10 patients without mitral disease served as the control group. MAD severity was measured from LA wall-posterior MV leaflet junction to the top of the LV infero-basal wall during end systole. Insterstitial fibrosis was assessed by calculating the extracellular volume (ECV) from T1 mapping of the left ventricular basal slice acquired before and after Gadolinium injection. Results Mean age was 47,5+\-23,3 years and 60% were male. ECV was higher in patients with MVP compared with controls (basal septum: 0.27 ± 0.04 vs 0.23 ± 0.03 p = 0.006; basal inferoposterior wall 0.28 ± 0.03 vs 0.23 ± 0.02 p = 0.003) and there was a significant correlation between MAD severity and ECV of the basal inferior wall (spearman rho 0.68, p &amp;lt; 0.0001) (Figure 1). Among MVP patients, ECV of the basal inferoposterior wall was higher in patients positive for LGE in the papillary muscles (ECV 0.31 ± 0.03 vs 0.27 ± 0.03 p 0,004). Conclusion In MVP patients, MAD severity was associated with a higher amount of interstitial fibrosis even in the absence of detectable macroscopic fibrosis in the papillary muscle region. Abstract 1047 Figure 1